Bug 32987 – Add build option ENABLE_XHTMLMP to Mac, GTK, and Chromium build systems

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Daniel Bates]

We should add the build option to enable XHTML-MP support to the Mac, GTK, and Chromium build systems (disabled by default).

Currently, passing the command-line argument --enable-xhtmlmp to build-webkit is only honored by the Qt and Windows builds.

[Daniel Bates]

Created attachment 45815 [details]
Patch

[WebKit Review Bot]

style-queue ran check-webkit-style on attachment 45815 [details] without any errors.

[WebKit Review Bot]

Attachment 45815 [details] did not build on chromium:
Build output: http://webkit-commit-queue.appspot.com/results/161668

[Daniel Bates]

CC'ing people knowledgeable about the various build systems, including Csaba Osztrogonac, Jan Alonzo, Eric Seidel, Adam Roben, and Mark Rowe

[Daniel Bates]

Created attachment 45817 [details]
Patch

Added missing comma after 'ENABLE_XHTMLMP=0' in WebKit/chromium/features.gypi.

[WebKit Review Bot]

style-queue ran check-webkit-style on attachment 45817 [details] without any errors.

[Csaba Osztrogonác]

WebKitTools/Scripts/build-webkit script pass to Qt build system if XHTML-MP is enabled or disabled. Now it is disabled by default, you can enable explicitly with: WebKitTools/Scripts/build-webkit --xhtmlmp

If you would like to use different default value for Qt port, you should put one of this line into WebCore.pro

!contains(DEFINES, ENABLE_XHTMLMP=.): DEFINES += ENABLE_XHTMLMP=0
or
!contains(DEFINES, ENABLE_XHTMLMP=.): DEFINES += ENABLE_XHTMLMP=1

Irrespectively of this modification, you can use --xhtmlmp and --no-xhtmlmp to toggle XHTML-MP support.

[Eric Seidel (no email)]

Comment on attachment 45817 [details]
Patch

Assuming HTMLNoScriptElement is appropriately guarded, this looks fine.

configure.ac looks insanely verbose and redundant.

Your ChangeLogs need lovin before you can land this:
+        Need a short description and bug URL (OOPS!)
+
+        No new tests. (OOPS!)
+

[Daniel Bates]

(In reply to comment #8)
> (From update of attachment 45817 [details])
> Assuming HTMLNoScriptElement is appropriately guarded, this looks fine.
> 
> configure.ac looks insanely verbose and redundant.

We should probably look to clean this up in a separate bug.

> 
> Your ChangeLogs need lovin before you can land this:
> +        Need a short description and bug URL (OOPS!)
> +
> +        No new tests. (OOPS!)
> +

Will change.

[Daniel Bates]

Committed r52972: <http://trac.webkit.org/changeset/52972>